This PR was submitted for the master branch but it was merged into the 2.7 branch instead (closes#17486).
Discussion
----------
[FrameworkBundle] Throw for missing container extensions
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | none
| License | MIT
| Doc PR | none
This covers the case when an existing bundle does not have an extension, and its config is dumped. Before, calling `app/console config:dump` on such bundle lead to FatalErrorException with ```Call to a member function getAlias() on null```, now we process such cases and throw an exception with some explanatory text.
Commits
-------
884368e Throw for missing container extensions
This PR was merged into the 2.8 branch.
Discussion
----------
Overriding profiler position in CSS breaks JS positioning
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
The positioning of a profiler info block (open to the left or right) is [calculated using Javascript](https://github.com/symfony/symfony/blob/master/src/Symfony/Bundle/WebProfilerBundle/Resources/views/Profiler/toolbar_js.html.twig#L35). Since Symfony 2.8, the config/version panel is right-aligned and opens to the left. If another panel is added to the right of it, the panel cannot open correctly.
Styles are unset in https://github.com/symfony/symfony/blob/master/src/Symfony/Bundle/WebProfilerBundle/Resources/views/Profiler/toolbar_js.html.twig#L46-L47 but that means it is set back to the stylesheet settings, which results in `right:0, left:0` on the element.
Manual testing is fairly easy: Just add a CSS class `sf-toolbar-block-right` on one or multiple panels (e.g. Doctrine) that result in the Config panel to have enough room to open to the right.
Here's a screenshot of the problem:
The other option would be to set the position in javascript to `right: auto` instead of unsetting, but I prefer to fix invalid CSS ;-)
Commits
-------
79474a6 Profiler CSS position conflicts with JS detection
This PR was submitted for the 2.8 branch but it was merged into the 2.7 branch instead (closes#16859).
Discussion
----------
[TwigBridge] add missing unit tests (AppVariable)
Just add few missing unit tests.
---
| Q | A
| ------------- | ---
| Bug fix? | [no]
| New feature? | [no]
| BC breaks? | [no]
| Deprecations? | [no]
| Tests pass? | [yes]
| Fixed tickets | [n/a]
| License | MIT
| Doc PR | [n/a]
Commits
-------
156cdb5 [TwigBridge] add missing unit tests (AppVariable)
This PR was submitted for the 2.8 branch but it was merged into the 2.3 branch instead (closes#16873).
Discussion
----------
Able to load big xml files with DomCrawler
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR |
Commits
-------
3dae825 Able to load big xml files with DomCrawler
This PR was squashed before being merged into the 2.3 branch (closes#16897).
Discussion
----------
[Form] Fix constraints could be null if not set
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
If the form options has no key for constraints the default should be an empty array to be ignored by the loop in the FormValidator. The default without this fix is ```null``` and foreach will throw an error.
The "Bug" also still exists in master-Branch.
Commits
-------
f80e0eb [Form] Fix constraints could be null if not set
This PR was merged into the 2.7 branch.
Discussion
----------
[Translation][Writer] avoid calling setBackup if the dumper is not FileDumper
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | ~
| License | MIT
| Doc PR | n/a
Commits
-------
369b2d4 [Translation][Writer] avoid calling setBackup if the dumper is not an instance of FileDumper.
This PR was merged into the 2.7 branch.
Discussion
----------
[FrameworkBundle] Compute the kernel root hash only one time
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | n/a
| License | MIT
| Doc PR | n/a
Commits
-------
23431e9 [FrameworkBundle] Compute the kernel root hash only one time
This PR was squashed before being merged into the 2.3 branch (closes#17134).
Discussion
----------
[Finder] Check PHP version before applying a workaround for a PHP bug
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | --
| License | MIT
| Doc PR | --
Commits
-------
cba206a [Finder] Check PHP version before applying a workaround for a PHP bug
This PR was merged into the 2.3 branch.
Discussion
----------
sort bundles in config:dump-reference command
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR |
This backports #17495 to the `2.3` branch.
Commits
-------
a5c8811 sort bundles in config:dump-reference command
This PR was submitted for the master branch but it was merged into the 3.0 branch instead (closes#17506).
Discussion
----------
[FrameworkBundle] enable assets when templates are enabled
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #16570
| License | MIT
| Doc PR |
This puts back the default behavior of Symfony 2.8 where the Asset
component features were implicitly configured with sensible default
values when no explicit configuration was present but the templating
section was enabled.
Commits
-------
d0de425 enable assets when templates are enabled
This puts back the default behavior of Symfony 2.8 where the Asset
component features were implicitly configured with sensible default
values when no explicit configuration was present but the templating
section was enabled.
This PR was merged into the 2.3 branch.
Discussion
----------
[2.3] [CS] MethodSeparationFixer - findings.
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| License | MIT
Th goal of this PR is _not_ to have it merged, but to see if the `MethodSeparationFixer` (https://github.com/FriendsOfPHP/PHP-CS-Fixer/pull/1426) is suitable for SF level/configuration in the PHP-CS-Fixer (and thereby might later become part of `fabbot.io`).
Commits
-------
d7b730f Fixer findings.
This PR was merged into the 2.7 branch.
Discussion
----------
[2.7][Asset] Add defaultNull to version configuration
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #14832, #16511
| License | MIT
| Doc PR |
In #16511 PR was omitted defaultNull version for one case.
Commits
-------
65adb72 add defaultNull to version
* 2.8:
Ability to set empty version strategy in packages
Display Ajax request from newest to oldest in the toolbar
CLI: use request context to generate absolute URLs
[SecurityBundle] Optimize dependency injection tests
Sort bundles in config commands
[HttpFoundation] Do not overwrite the Authorization header if it is already set
tag for dumped PHP objects must be a local one
* 2.7:
Ability to set empty version strategy in packages
CLI: use request context to generate absolute URLs
[SecurityBundle] Optimize dependency injection tests
Sort bundles in config commands
[HttpFoundation] Do not overwrite the Authorization header if it is already set
tag for dumped PHP objects must be a local one
* 2.3:
[SecurityBundle] Optimize dependency injection tests
[HttpFoundation] Do not overwrite the Authorization header if it is already set
tag for dumped PHP objects must be a local one
This PR was submitted for the master branch but it was merged into the 3.0 branch instead (closes#17508).
Discussion
----------
Fix repeated "has been" in two sentences of UPGRADE-3.0 file
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | N/A
| License | MIT
| Doc PR | N/A
In file UPGRADE-3.0 `has been` is repeated in two sentences.
Commits
-------
d83c92c Fix repeated has been in two sentences
This PR was merged into the 2.7 branch.
Discussion
----------
[2.7][Asset] Ability to set empty version strategy in packages
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no, but not sure, as for me test was wrong
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #14832
| License | MIT
| Doc PR |
Comment about '' thing. As you can see in xml test - we can't for attribute set null value. And for xml version '' empty string considered as null value, that affect all others formats and test is changed.
Commits
-------
646fc9c Ability to set empty version strategy in packages
This PR was merged into the 2.7 branch.
Discussion
----------
[2.7] [FrameworkBundle] [DX] Sort bundles in config commands
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | none
| License | MIT
| Doc PR | none
Reopened against `2.7` instead of #17490
This makes the behaviour of config commands a bit more user-friendly. Scanning through an ordered list should be a lot easier for the user, especially if she knows what's the exact bundle they're looking for. The sort is applied to the bundle name, as the extension alias might not always be present.
Before:
After:
Commits
-------
80d51e0 Sort bundles in config commands
This PR was merged into the 2.7 branch.
Discussion
----------
[Asset] CLI: use request context to generate absolute URLs
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #15448
| License | MIT
| Doc PR |
Commits
-------
766a648 CLI: use request context to generate absolute URLs
This PR was squashed before being merged into the 2.3 branch (closes#17497).
Discussion
----------
[SecurityBundle] Optimize dependency injection tests
| Q | A
| ------------- | ---
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #17366
| License | MIT
| Doc PR | -
Tests building the same container multiple times. The same solution can be applied as here: #17399. This time I ran comparison tests only on 2.3. (my laptop is really slow)
**2.3 before**
```go
Time: 14.79 seconds, Memory: 48.25Mb
OK (77 tests, 227 assertions)
OK src/Symfony/Bundle/SecurityBundle/
```
**2.3 after**
```go
Time: 9.78 seconds, Memory: 48.75Mb
OK (77 tests, 227 assertions)
OK src/Symfony/Bundle/SecurityBundle/
```
Commits
-------
bf56d2f [SecurityBundle] Optimize dependency injection tests
This PR was submitted for the master branch but it was merged into the 2.3 branch instead (closes#17478).
Discussion
----------
[HttpFoundation] Do not overwrite the Authorization header if it is already set
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #17345
| License | MIT
| Doc PR | -
Commits
-------
53ebfda [HttpFoundation] Do not overwrite the Authorization header if it is already set
This PR was merged into the 2.3 branch.
Discussion
----------
[Yaml] tag for dumped PHP objects must be a local one
| Q | A
| ------------- | ---
| Bug fix? | yes
| New feature? | no
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets | #16877
| License | MIT
| Doc PR |
Commits
-------
9d03478 tag for dumped PHP objects must be a local one
